Overwhelmingly, it sounds like I would never want any of you handling financial decissions at any company I am eployed with.
Capital expenditure, risk assesment, potential/probable return. Taking a look at all of these factors- why would a company in nV's position want to even think about signing a deal with MS? With a royalty kickback model in place if we assume that MS is going to double their sales over last generation, a task that would never be assumed in a reasonable risk assesment model, then they are looking at a potential of 50Million units. With only royalty fees being paid on a per chip basis the idea of diverting any reasonable amount of engineering resources to the project made next to no sense. A heavily customized design with only a partial ability to cross utilize the technology developed- it isn't a winning combination for a company that was in nV's position.
This has almost nothing to do with the FX series nor does it have to do with the 6 series. Both of those teams were quite occupied in the timeframe that we are talking about. The team working on the G80 design or the G70 team would be the only viable choices- and either of those teams would have left rather sizeable gaps in their lineup. Diverting a sizeable portion of either team for the potential ROI for the 360 vs the potential ROI in the PC space didn't make any sense. From a business standpoint, using a relatively modified version of an off the shelf part with a realistic potential ROI of double to quadruple what the 360 offered is a no brainer(between contract and reasonable installed base expectations base on market trends).
The best way to think of nV's choices v ATi is that nV had a market cap of nearly double ATi prior to the acquisition announcement. It isn't about designing the 'coolest' technology- it is about utilizing your resources to fill market demands in the most reasonable manner possible.
Obviously the argument will be put forth about the unified architecture paying long term dividends for ATi- the problem with that assumption is that we haven't seen anything to back that up yet. On paper the design theory behind the P4 should have wiped the floor with the Athlon. On paper most people on these forums thought that the FX was going to edge out the R300. Maybe it will pay off for them, I'm not saying it won't. What I am saying is that with their efforst focused on unified they don't have the expertise in knowing what would have been possible with segmented that nVidia has. That works both ways without a doubt, but most seem to be assuming that US is somehow in and of itself going to give ATi a big edge when we have absolutely nothing to back that assertion up yet. Perhaps it will and ATi's bet will have paid off. nV's did for the original XBox. For all the talk about their shortcomings of the FX series, I think if you gave nV the option that they would lose out for a single generation in the GPU segment and become the second largest chipset manufacturer in the world to go along with that they would have taken the deal. It may end up being a long term stroke of genius for ATi, but it is a much larger gamble then what nV did for Sony. Maybe the ROI will be huge, but the downside could end up being comparable(although now it will largely be masked either way with the acquisition).
